Historic Learning Approach for Auto-tuning OpenACC Accelerated Scientific Applications

Handle URI:
http://hdl.handle.net/10754/622145
Title:
Historic Learning Approach for Auto-tuning OpenACC Accelerated Scientific Applications
Authors:
Siddiqui, Shahzeb; Alzayer, Fatemah ( 0000-0003-4393-1457 ) ; Feki, Saber
Abstract:
The performance optimization of scientific applications usually requires an in-depth knowledge of the hardware and software. A performance tuning mechanism is suggested to automatically tune OpenACC parameters to adapt to the execution environment on a given system. A historic learning based methodology is suggested to prune the parameter search space for a more efficient auto-tuning process. This approach is applied to tune the OpenACC gang and vector clauses for a better mapping of the compute kernels onto the underlying architecture. Our experiments show a significant performance improvement against the default compiler parameters and drastic reduction in tuning time compared to a brute force search-based approach.
KAUST Department:
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division; Extreme Computing Research Center; KAUST Supercomputing Laboratory (KSL)
Citation:
Siddiqui S, AlZayer F, Feki S (2015) Historic Learning Approach for Auto-tuning OpenACC Accelerated Scientific Applications. High Performance Computing for Computational Science -- VECPAR 2014: 224–235. Available: http://dx.doi.org/10.1007/978-3-319-17353-5_19.
Publisher:
Springer Science + Business Media
Journal:
High Performance Computing for Computational Science -- VECPAR 2014
Conference/Event name:
11th International Conference on High Performance Computing for Computational Science, VECPAR 2014
Issue Date:
17-Apr-2015
DOI:
10.1007/978-3-319-17353-5_19
Type:
Conference Paper
ISSN:
0302-9743; 1611-3349
Appears in Collections:
Conference Papers; KAUST Supercomputing Laboratory (KSL); Extreme Computing Research Center; Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division

Full metadata record

DC FieldValue Language
dc.contributor.authorSiddiqui, Shahzeben
dc.contributor.authorAlzayer, Fatemahen
dc.contributor.authorFeki, Saberen
dc.date.accessioned2017-01-02T08:10:21Z-
dc.date.available2017-01-02T08:10:21Z-
dc.date.issued2015-04-17en
dc.identifier.citationSiddiqui S, AlZayer F, Feki S (2015) Historic Learning Approach for Auto-tuning OpenACC Accelerated Scientific Applications. High Performance Computing for Computational Science -- VECPAR 2014: 224–235. Available: http://dx.doi.org/10.1007/978-3-319-17353-5_19.en
dc.identifier.issn0302-9743en
dc.identifier.issn1611-3349en
dc.identifier.doi10.1007/978-3-319-17353-5_19en
dc.identifier.urihttp://hdl.handle.net/10754/622145-
dc.description.abstractThe performance optimization of scientific applications usually requires an in-depth knowledge of the hardware and software. A performance tuning mechanism is suggested to automatically tune OpenACC parameters to adapt to the execution environment on a given system. A historic learning based methodology is suggested to prune the parameter search space for a more efficient auto-tuning process. This approach is applied to tune the OpenACC gang and vector clauses for a better mapping of the compute kernels onto the underlying architecture. Our experiments show a significant performance improvement against the default compiler parameters and drastic reduction in tuning time compared to a brute force search-based approach.en
dc.publisherSpringer Science + Business Mediaen
dc.titleHistoric Learning Approach for Auto-tuning OpenACC Accelerated Scientific Applicationsen
dc.typeConference Paperen
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Divisionen
dc.contributor.departmentExtreme Computing Research Centeren
dc.contributor.departmentKAUST Supercomputing Laboratory (KSL)en
dc.identifier.journalHigh Performance Computing for Computational Science -- VECPAR 2014en
dc.conference.date2014-06-30 to 2014-07-03en
dc.conference.name11th International Conference on High Performance Computing for Computational Science, VECPAR 2014en
dc.conference.locationEugene, OR, USAen
kaust.authorSiddiqui, Shahzeben
kaust.authorAlzayer, Fatemahen
kaust.authorFeki, Saberen
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.